FNZ 68 - Simuliidae (Insecta: Diptera) - Biogeography
Craig DA, Craig REG, Crosby TK 2012. Simuliidae (Insecta: Diptera). Fauna of New Zealand 68, 336 pages.
(
ISSN 0111-5383; no. 68 (print),
ISSN 1179-7193 (online)
;
no.
68.
ISBN 978-0-478-34734-0 (print),
ISBN 978-0-478-34735-7 (online)
).
Published 29 June 2012
ZooBank: http://zoobank.org/References/9C478D54-FEB2-45E8-B61C-A3A06D4EB45D
Biogeography
Avise (2004b) in a brief overview of biogeography made the distinction between cladistic biogeography based on morphological information, that can deal with deeper temporal scales at high taxonomic levels, and phylogeography which uses molecular data and provides insights at the intraspecific level. Since both our morphological and molecular data are analysed using cladistics, we simply use the term biogeography.
A major question that faces New Zealand biogeographers is the origin of New Zealand’s biota. Is it of ancient Gondwana mien (vicariant origin) or of more recent origin via dispersal? Central to this question is the suggested Oligocene Inundation (e.g., Landis et al. 2008), when some assert that New Zealand was completely submerged ca 34–23 Mya. This has been hotly debated (e.g., Cooper & Cooper 1995; Waters & Craw 2006; Wallis & Trewick 2009; Giribet & Boyer 2010; Tennyson 2010) and there is fossil evidence that shows there was a considerable vertebrate fauna present in southern New Zealand in the Miocene immediately following the Oligocene, indicative of extensive land (Worthy et al. 2006; Worthy et al. 2011). High conifer diversity in the Oligo-Miocene of New Zealand also supports considerable land being present at that time (Jordan et al. 2011), so complete inundation does not appear to have occurred.
For organisms lacking fossil evidence, one possible means of addressing origins is to examine relationships of biota to antecedents, namely from Australia and South America. A famous example of this is Brundin’s (1966) study of podonomine midges (Chironomidae). As reiterated by Gibbs (2006) and others, closer relationships of New Zealand’s biota to that of South America than to that of Australia, indicate older vicariant origins that trace back to Gondwana. Closer relationship with Australia than South America suggest more recent dispersal from Australia is likely, while not disproving earlier Gondwanan origins.
A critical part of such examinations now involves determining rates of genetic divergence between the clades of organisms of interest — the so-called “molecular clock” (e.g., Brower 1994; Gaunt & Miles 2002; Trewick & Morgan-Richards 2005; Wallis & Trewick 2009). An estimated time of divergence since the Oligocene Inundation would indicate subsequent dispersal to New Zealand. Determining time of divergence between lineages can involve invocation of significant geological events, such as the break-up of Gondwana and the splitting of Zealandia from Gondwana at some 83 Mya. For finer analysis, more recent events such as formation of the Southern Alps (ca 5–3 Mya), to the formation of the modern Cook Strait (ca 0.45 Mya), or climatic events such as glaciations can be used. Or, at times, a general rate of divergence (% change in molecular genetic bases per million years) can be applied; however, such use is problematic since rates of change may differ between lineages, and there are other caveats, some rather firmly put (Avise 2004a; Rubinoff 2006; Rubinoff et al. 2006).
We use both general rate of divergence and time of geological events in discussing the age of Austrosimulium and its intrageneric clades.
Paleological aspects and effects on New Zealand
Geological
Much has been written about the geological underpinnings of New Zealand and its rather tortuous history. The brief overview below is based, in large part, on publications by Fleming (1975, 1979), Suggate et al. (1978), Stevens (1980), Coates (2002), Gibbs (2006), Campbell & Hutching (2007), Graham (2008), and many other specific studies (e.g., Cook et al. 1999; King 2000; McLoughlin 2001, Trewick & Bland 2011). The more recent publications should be consulted for entry into the extensive literature.
Evidence derived from this study indicates that New Zealand Austrosimulium is at maximum 8 My old, and more likely 5 My old. For the paleogelogical aspects of New Zealand relevant to Austrosimulium we deal briefly with the Oligocene, but mainly from the orogeny of the Southern Alps onwards. A review of earlier events involved in the geological evolution of New Zealand, from Gondwana to the Oligocene is available online in the Supplementary Material at <fnz.landcareresearch.co.nz>.
At the end of the Oligocene (25 Mya), the Pacific Plate began to collide with and subduct beneath the Australian Plate along the north of New Zealand. This was the beginning of the Kaikoura Orogeny and a transform fault — the incipient Alpine Fault. At that time the Fault ran more or less east–west. By the late Miocene (10 Mya) that subduction had begun south of the New Zealand landmass. These events meant that movements along the Alpine Fault were sideways: currently shown by the 460 km that separate similar old geological formations in Nelson and Southland. By later Miocene and early Pliocene (8–5 Mya) the Alpine Fault began to rotate anticlockwise. The Pacific Plate, overlain with greywacke rock, moved sideways along the Australian Plate (comprised largely of granite), and began to push west against the latter leading to uplift of the Southern Alps. This is the reason for the unusual geological arrangement in the Southern Alps of granite to the west and various greywackes to the east. There is good evidence that the major orogeny of the Southern Alps started in the early Pliocene at 5 Mya (Chamberlain et al. 1999; Chamberlain & Poage 2000; Batt et al. 2000). Older dates are now thought to be incorrect (Shuster et al. 2011).
At that time in the North Island there was a Manukau Strait separating Northland from the rest, and the Waitakere Ranges were probably islands. Trewick & Bland (2011), however, in an examination of the geological relationships between the North and South Islands over the last 4 My, noted that previous ideas regarding broad marine separation between the two islands needed to be reconsidered. Cook Strait, as presently known, was not in existence 4 Mya; instead to the west the 2 islands were separated by a deep Wanganui Basin, and to the east by a narrow so-called Kuripapango Strait across a finger of high land extending from the South Island. There was no broad Manawatu Strait as often suggested. The South Island extension from what is now the Marlborough Sounds went northeast to about present-day Palmerston North, and has been termed the “Wanganui–Marlborough Shield”. By 3 Mya (Late Pliocene) the rates of vertical tectonism had increased and what was the Wanganui Basin to the west began to rise. The Kuripapango Strait to the north closed, and farther south was a narrower Manawatu Strait that passed through the South Island extension. There is evidence for a northeast band of higher land, separated from the proto-North Island by the Ruataniwha Strait. The Wanganui–Marlborough Shield was fragmenting, and by 2.4 Mya (Early Pleistocene) there was a broad, shallow Manawatu Strait, but it appears as if it was constrained to the west and east by extremely shallow water or low land. Farther east there appears to have been a range of shoals and islands, separated from the two islands by the Ruataniwha Strait. At 1 Mya the present-day North and South Islands were close to modern configuration, and the Manawatu Strait was obliterated by the orogeny of modern south-central mountain ranges. There was a narrow land connection across the present Cook Strait. Marlborough Sounds were well foundered and near to modern conformation, and what would become the Tararua Range north of Wellington was rising. With continued land movements, the modern Cook Strait was punched through at about 0.45 Mya during a mid Middle Pleistocene interglacial. During the Waimaunga glaciation (0.30 Mya) the North Island and South Island were connected by a low, gravel-based, Marlborough–Taranaki Plain. Rivers from Taranaki, namely the Rangitikei and Wanganui, flowed out onto this plain. From the South Island the Waimea River possibly flowed northeast, but did not merge with those from the north. The plain was submerged during the late Pleistocene interglacial (0.10 Mya). During the Last Glacial Maximum (0.02 Mya) and with sea level depressed as much as 135 m, the two islands are generally considered to have again joined. However, with intertidal scouring and tectonic movements, Cook Strait was deeper and the junction narrower and well to the west — the Farewell Rise. A junction is disputed by some. Regardless, if there was a channel, or channels, across the plain, they were narrow, and the connection of land, if any, transitory (see below, p. 70; Fig. 515–518).
Volcanism
Intraplate volcanics have been important in formation of New Zealand’s landscape. Subduction zones between plates normally have associated volcanic activity along the interaction zone, however, distant from the junction. This derives from the subducted plate as it is heated and rises as magma, forming a volcanic arc. Such volcanic activity began in the north of New Zealand in the Early Miocene (ca 23 Mya) producing volcanoes along Northland. By Middle Miocene (ca 20–18 Mya) the volcanism had moved to along the present-day Coromandel Peninsula and continued in various forms until the end of the Miocene (ca 6 Mya). In Early Pliocene times (ca 5–2 Mya) the volcanism had moved farther east to near the present-day Kaimai Range and Tauranga, and at 1.9–1.6 Mya (Briggs et al. 2005) during a fundamental rearrangement of plate movement in the New Zealand region activity switched to the Taupo Volcanic Zone (TVZ; sometimes referred to as the Central Volcanic Region). There are, however, estimates that volcanism might have started there around 3 Mya.
Vast ignimbrite eruptions associated with the currently-active Taupo Volcanic Zone continued into the early Quaternary (1 800 years BP). Some of the pyroclastic flows associated with these eruptions are considered to be amongst the largest on Earth. One, at approximately 1.0 Mya, which in part produced the (Cape) Kidnappers Ignimbrite, extended as far north as Auckland. Volcanism associated with what is now Lake Taupo has coated extensive areas of the North Island with ash and pumice many times. One pyroclastic flow at 22 000 years BP, for example, covered north to Whakatane, east to Napier and Gisborne, south to Wellington, and west to Mount Taranaki.
The three volcanic mountains at the centre of the Volcanic Plateau are relatively young and represent the southernmost extent of the Taupo Volcanic Zone. Tongariro to the north is the oldest, considered to have commenced erupting around 275 000 years BP, reaching its modern form about 65 000 years BP. Ngauruhoe, in the middle, is judged by some to be merely 2 500 years old; however, it has been continuously active, resulting in its mass. Ruapehu is thought to have started erupting about 245 000 years BP. The Volcanic Plateau, as presently known and some 500 m above the surrounding landscape in places, is a product of those volcanoes and relatively recent in its present form.
Volcanism in the Taranaki region is the westernmost extent of the northern intraplate volcanism and began in the Late Pliocene (ca 1.8 Mya). The iconic Mount Taranaki (0.015–0.012 Ma) and its older precursors Mounts Pouakai (0.25 Mya), Kaitake (0.52 Mya), and the Sugar Loaf Islands (1.75 Mya) to the west are derived from deeper parts of the subducting Pacific Plate.
Formation of the volcanic Banks Peninsula on the east coast of the South Island is of relevance here because of the precinctive aquatic fauna (Winterbourn 2008), but not for simuliids. Volcanism of the peninsula began in the Middle Miocene (15 Mya) and ceased about 5-6 Mya.
The geological underpinnings of Stewart Island to the south of the South Island are ancient. Simply the island consists of two main units; the southern Fiordland Terrane (Mid Cambrian, 510 Mya), and the northern Brook Street Terrane (Permian, 265 Mya), separated by a major fault line — the Median Tectonic Line of some — currently occupied by the Freshwater River. That the island has been above sea level for a long time is illustrated by the exfoliated, rounded, granite massifs of Gog and Magog (Cretaceous, 125–105 Mya) on the southern Fiordland Terrane. There is little evidence of glaciation on the island, its overall altitude being too low (McGlone & Wilson 1996). Foveaux Strait between Stewart Island and the South Island is shallow and was exposed during glaciation maxima sea level depressions (Fig. 518). Earlier, the Strait was likely to have been deeper, having been filled with outwash gravels during glaciations.
New Zealand is geologically a restless place! The rate of uplift of the southeastern flank of the North Island, and that of the main axis of the South Island is startling, ranging from 1–10 mm/year and has had major effects on the configuration of Cook Strait in recent geological time. Slip movement along the Alpine Fault Zone and the Southern Alps is estimated at 40 mm a year in some places.
To the east of South Island are the Chatham Islands. Simuliids are not known from there, even though there is suitable running water, so we do not comment further.
Of more interest are the subantarctic islands of the Campbell Plateau, in particular the Auckland Islands and Campbell Island from which simuliids are known. These island groups are considered to be intraplate hot-spot shield volcanoes where magma penetrated the granite basement and sedimentary rocks of the Campbell Plateau. The Auckland Islands were formed from two volcanoes, Carnley and Ross, during the Middle Miocene (ca 19–12 Mya). In the Middle to Late Miocene (ca 11–6.5 Mya, mainly ca 7.0 Mya), volcanism shifted, with plate movements, to form Campbell Island. Earlier rocks found on these islands are considered by some to have been brought up from the underlying Campbell Plateau; such contaminants are common to intraplate volcanism.
Paleoclimate
Superimposed on the rapidly changing geology of New Zealand has been a series of glaciations — some 9–10 in the last million years (Burrows 2005). These are of relevance for a variety of reasons, but in particular to Simuliidae because of their requirement for running water. Apart from their biogeographic implications (e.g., McCulloch et al. 2010), New Zealand cold periods also serve as a test for models of glaciation elsewhere in the world, such as Antarctica and the Northern Hemisphere (e.g., Carter & Gammon 2004; Alloway et al. 2007; Schaefer et al. 2009). Because of the impact of glaciation on the landscape, ice ages are often considered along with the paleogeology of New Zealand (e.g. Shuster et al. 2011 and other citations in paleogeology section above). Environmental changes from the beginning of the Quaternary until Present have been reviewed by Newnham et al. (1999) and a detailed examination of glaciations and extent of glaciation for the South Island has been made by Burrows (2005); the following is based largely on these works.
Development of the Antarctic Circumpolar Current and cooling of Antarctica began at the upper end of the Eocene (34 Mya) with the opening of the Drake Passage between South America and Antarctica, and the Tasman “gateway” between Antarctica and Australia (Katz et al. 2011).
Dating glaciations can be difficult, since successive glaciation events obliterate earlier episodes. Therefore, dating is often done by examination of off-coast marine sediment cores, determining gamma ray emissions and oxygen isotopes, amongst other methods. The oxygen isotopes are proven proxies for temperature and, indirectly, sea levels. High sedimentation rates indicate glaciation on shore. Quartz grains in offshore sediments are indicative of strong winds, dry conditions, and loess formation (e.g., Nelson et al. 1993; Carter & Gammon 2004). Such signals are grouped into Marine Isotope Stages (MIS) and can have a resolution of around 100 years, although older records are at resolution of 1 000–2 000 years. For New Zealand these have been determined back to ca 4 Mya.
The earliest direct evidence for glaciation in New Zealand is that for the Ross Glaciation perhaps 2.55–2.45 Mya (end of the Pliocene), followed later by that of the Porika 2.15–2.10 Mya (beginning of the Pleistocene). Then there was a long period up to ca 900 000 years BP with some 20 alternating cool and warm periods. A period of some 9 or 10 regular glaciations (40 000–50 000 year periodicity) followed, for which there is better evidence. The last five of these glaciations, progressively, are named the Kawhaka, Nemona, Waimaunga, Waimea, and Otira Glaciations. The Otira appears to have been one of the great glaciations and was ca 75 000–14 500 years BP. Temperatures are known to have fluctuated during this time and glaciers were probably at their maximum extent at 70 000 years BP. At 26 500 years BP the Taupo caldera in the North Island erupted and produced a widespread Kawakawa (Oruanui) Tephra that allows definitive dating. There were even vaster eruptions earlier (ca 0.34 Mya), such as that of the Rangitawa (Holt et al. 2010), and such eruptions have been suggested as tipping points contributing to glaciations. More recent eruptions have also proven useful for dating (Froggatt & Lowe 1990; Newnham et al. 2003; Manville & Wilson 2004).
During the Otira Glaciation there were brief interstadial or warmer periods. There were two distinct ice advances: one at ca 65 000 years BP (MIS 4) and the other more recently, at ca 18 000 years BP (MIS 2c) — the former being slightly more extensive (McCarthy et al. 2008). There is evidence that at maximum advance the Rakaia Glacier (Burrows 2005; Alloway et al. 2007) extended as far east as present-day Methven. Similarly, the Rangitata Glacier to the south extended to near Mayfield. Burrows (2005) and others (e.g., Soons 1994) showed that at their final advance the eastern glaciers (Waimakariri, Rakaia, and Rangitata) terminated upstream of their present-day gorges. Farther south in the Mackenzie Basin ice extended beyond the present southern ends of Lakes Tekapo, Pukaki, and Ohau, and other lakes (Burrows 2005). There is evidence that during the Waimaunga Glaciation (ending 220 000 years BP) the Waimakariri, Rakaia, and Rangitata glaciers extended even farther eastwards than during the Otira Glaciation (Burrows 2005). These glacier advances are of relevance to refugial distributions of Austrosimulium.
Relevant as well to this study are the drops in sea level and shifts of climatic zones during glacial maxima. Sea level depression during the Last Glacial Maximum is generally cited at -130 m (Thomas et al. 2009), although other work (e.g., Thompson & Goldstein 2006) suggested -135 m. This means that a considerable area of the New Zealand continental shelf was exposed. Reconstructions of land for that time usually show only that for the maximum sea level depression (e.g. Burrows 2005; Alloway et al. 2007). They show Stewart Island as broadly connected to the South Island (e.g. Fig. 518), but not extending far south. Foveaux Strait at present is shallow, ranging from 20–50 m in depth over much of its extent. Exposed land continued up the east coast through Dunedin and as an extensive Canterbury Plain, extended some 40 km farther east than present coastlines (Alloway et al. 2007). Banks Volcanoes were well connected to the Plain, as they were during other sea level depressions. That connection became permanent in late Pleistocene times (ca 75 000 years BP) (Shulmeister et al. 1999), in large part with gravel from the Southern Alps, transported and deposited by the Waimakariri River. In more recent times (e.g., 6 500 years BP), that connection has, however, occasionally been narrowed (Burrows 2005). Along the West Coast where there is less continental shelf, ice extended out to sea along Fiordland, while farther north there were small pockets of land between glaciers that extended to the sea; well illustrated by Burrows (2005). Northwest Nelson and Taranaki were at times connected via the Farewell Rise, but how well is debatable; this is discussed on p. 70. Around the North Island there was extensive land up the west coast. Many of the present smaller offshore islands, particularly to the northeast, were connected to the mainland.
Stewart Island, because of its relatively low altitude and marine influence had little glaciation (McGlone & Wilson 1996; Alloway et al. 2007), but the main lineage of the Southern Alps and the Kaikoura Ranges were heavily glaciated and many of the current landforms observed are a result of that, as are distributions suggested for some biota (e.g., McCulloch et al. 2010; and this study). In the North Island there was minor glaciation on the volcanoes central to the Volcanic Plateau and, to the west, on Mount Taranaki.
There was extensive modification to the climate and hence vegetation; both have been investigated in detail (e.g., Burrows 2005; Alloway et al. 2007). This is of relevance to distribution of A. ungulatum, that appears to rely in part on forest. Other species have particular temperature requirements.
During the Last Glacial Maximum (MIS 2) there is evidence that extremely cold air derived from Antarctica reached New Zealand along with icebergs. The “equilibrium line” (snowline) has been determined for various periods and, for example, at the middle of the Otira Glaciation, some 45 000 years BP, it was 850–1 300 m below current levels, with the mean temperature depressed by approximately 4.5–7.0 C°, with concomitant downward movement of vegetation zones (e.g., Burrows 2005; Marra 2006; Alloway et al. 2007).
For Stewart Island, pollen analyses by McGlone & Wilson (1996) showed a hardwood forest and many tree ferns were present prior to 9 000 years BP and up to 5 500–4 500 years BP; then there were major changes. Absence of some tree species, including the iconic southern beech, Nothofagus, is attributed to failure to disperse plus limited time for which suitable habitats have been available.
Much of interior Southland and the Catlins region in the South Island appear to have had extensive regions of shrub, tussock, herb, and fell fields, and to have been above the treeline. These extended north, narrowed opposite Banks Peninsula, to expand again around the Kaikoura Ranges and northwest into the Nelson area. Major glaciers, in particular the Rakaia and Rangitata, at their maxima, punched through this zone (Burrows 2005; Alloway et al. 2007). At lower altitudes there were extensive tracts of dry, open grassland with some shrubs, ranging from Stewart Island, north throughout the extended Canterbury Plains, and up to Kaikoura. The distribution of plants in the Nelson–Marlborough area indicates presence of refugia (e.g., Marra et al. 2009).
Of particular interest to this study of Austrosimulium is why some South Island species do not occur on the North Island. This question too was of particular interest to Dumbleton (1973). Cook Strait is an obvious barrier to dispersal, but its present configuration is relatively modern. Using bathymetric data of the Greater Cook Strait (Lewis et al. 1994; Gillespie & Nelson 1996; and that from the Ministry for Primary Industries, National Aquatic Biodiversity Information Systems 2010) and high resolution sea level data for the last 250 000 years (Thompson & Goldstein 2006), the extent of the putative land bridge, the Farewell Rise, between Cape Farewell and Cape Egmont can be detailed over that period (Fig. 517). Similarly, this can be done for the Three Kings Islands (Fig. 516), and Stewart Island plus Foveaux Strait, and Auckland and Campbell Islands (Fig. 518). We assume the bathymetry in the Greater Cook Strait has not changed greatly over the period considered here; debatable given the rate at which sediments are laid down and scoured, and uplift is occurring (Lewis et al. 1994; Mountjoy et al. 2009).
For Cook Strait, a deep (-200 to -300 m) D’Urville Canyon currently extends northwest from the Cook Strait Narrows canyon to the middle of the Taranaki Bight. It shallows rapidly to the west, in narrow configuration, up to some -100 m, to approximately the same longitude as Mount Taranaki to the north. The deepest portion of the present Farewell Rise is ca -70 to -90 m. Depression of the sea to that level would have exposed extensive low relief land adjacent on the North and South Islands. But the westernmost extent of the D’Urville Canyon has a depth of -94 m, and depression of sea level greater than that would be needed, at present, to close the gap between the two islands.
Sea level reconstruction shows that between 250 000 and 185 000 years BP (MIS 7) levels varied between modern levels, down to -30 m. Stewart Island would probably have been partially connected to the South Island, perhaps a connection as wide as 50 km. There was then a rapid drop in level down to perhaps -70 m (detailed data are not available) with a rapid rebound to ca -45 m at 175 000 years BP until ca 165 000 years BP (early MIS 6, Waimea Glaciation). Foveaux Strait would have been well exposed off and on (50–70 km wide); Farewell Rise would not have been, although Cook Strait would have been reduced to some 70–80 km wide. Data are again sketchy between then and 135 000 years BP, and Thompson & Goldstein (2006) made no reconstruction of sea levels for that period. However, associated d18O isotope data suggested sea level depression was probably the equal of following major low sea stands (see interpolated levels from Huybrecht 2002). Therefore, portions of Farewell Rise could, off and on, have been exposed during that period, but not markedly, and probably never for more than ca 10 000 years. At 135 000 to130 000 years BP, sea level rose and by 125 000 years BP, was at modern high stands and even higher by some +10 m (end of MIS 6 and Waimea Glaciation). Any exposed land on the Farewell Rise was well covered, as was Foveaux Strait. Levels dropped again at ca 115 000 years BP, to -20 m (parts of Foveaux Strait may have been variously exposed) until ca 100 000 years BP, with a marked depression to ca -60 m, at 95 000 years BP (Foveaux Strait would have been exposed and 80–90 km wide), with levels then rising again to -20 m. Sea levels were variable over the Kaihinu interglacial between the Waimea and Otira Glaciations. Beginning at 75 000 years BP (start of MIS 4, 3, 2, and the Otira Glaciation) there was a major depression of sea level, down to some -75 m, with fluctuations, at which time the Farewell Rise might have been irregularly exposed, or with a markedly narrowed Cook Strait of ca 60 km, for short periods around 40 000 years BP. Levels continued dropping to -90 m at about 60 000 years BP at which time there could have been a full connection between the North and South Islands, but the connection would have been narrow, a mere 20–30 km wide. From then and 30 000 years BP, levels fluctuated between -55 m and -95 m, so the Rise would have been variously exposed and connected over a period of some 30 000 years. Then, commencing at that latter time, there was a precipitous drop in sea levels down to some -135 m at ca 20 500 years BP. This was immediately followed by a short-lived rise to -110 m, associated with the MIS 2b interstadial, then back down to the low stand. That deep low stand is associated with MIS 2 of the Last Glacial Maximum when there was an expansion of glaciers in New Zealand (e.g., McCarthy et al. 2008). From then until ca 17 500 years BP sea levels rose rapidly, reaching -90 m and less at ca 18 000 years BP, at which time the Farewell Rise was inundated again and has remained completely so since.
At maximum exposure Farewell Rise might have approached 150 km wide, but for perhaps only 5 000 years. This maximum exposure is that usually illustrated (e.g., Alloway et al. 2007; Trewick & Bland 2011) and tends to give the wrong impression of how much and how long land was exposed during the last glaciation. But for much of its exposed existence the Rise was not much wider than 20–50 km, and this, correctly, is illustrated by Lewis et al. (1994) and Marra et al. (2009). Further, given that much of the Rise was terrigenous gravel, plus fine sediment derived from the glaciers to the southwest, brought north by the Westland Current and passed to the D’Urville Current through Cook Strait, as well that weather systems were farther north during glacial periods and strength of westerly winds exacerbated (Fitzharris et al. 1992; Vandergoes & Fitzsimons 2003; Carter & Gammon 2004; Toggweiler 2009; and others), we suggest that the exposed Farewell Rise, particularly when narrow, was generally wind-driven drifts of fine material not conducive as a habitat to most organisms. This is in general agreement with Alloway et al. (2007) and others who suggested that the Rise was a shrubland–grassland mosaic at its greatest extent. At considerable variance with this is Marra et al. (2009), who were of the opinion that the junction was forested. Worthy & Holdaway (2002) were of the opinion that for now extinct birds, the moa, the junction was not of great dispersal significance during the Last Glacial Maximum. Indeed, they illustrated a narrow marine passage across Farewell Rise during that period (their Fig. 1.2). We do not greatly disagree. Newnham et al. (2003) considered that at about 18 000–16 000 years BP there was a marked reduction and southwards contraction of the strong westerly winds that characterised the Last Glacial Maximum climate, about the time the Farewell Rise was being flooded (Fig. 515).
The present bathymetry of western Cook Strait indicates that even if well exposed and connected — but not broad — Farewell Rise most likely consisted of narrow tidal scour channels separated by low rises, certainly at the -80 to -90 m sea level (superficially indicated in Fig. 517), which was the majority of the situation during the Otira Glaciation (Fig. 515). There is no question that modern tidal scour in Cook Strait is severe, in large part because of the offset times of tidal maxima between the east and west coast of the islands. We assume it was so during glaciations, perhaps exacerbated by the northerly shifts in climate.
Detailed bathymetry of Foveaux Strait (e.g., Cullen 1967) showed that at its present depths, regions around the easterly Ruapuke Island between the South Island and Stewart Island would have been close to connected at a depression of sea level of a mere -25 m, except perhaps for a narrowed channel a kilometre wide. At sea level depression of -40 m, connection to the South Island would have been well established to the east through Ruapuke Island and well to the west. At -50 m depression the connection could have been as wide as 50–60 km, but of low altitude and relief. At -100 m depression, the connection was in excess of 150 km wide.
Also of relevance to historical biogeography of New Zealand simuliids is the possibility, during sea level depression, that rivers from the main islands merged, or were in close proximity. For the sea level depression during the Mid Pleistocene glaciation at ca 1.25 Mya, Lewis et al. (1994: their Fig. 13e) illustrated rivers from the North Island (e.g., Wanganui and Rangitikei) as flowing onto the connecting plain to the South Island. These rivers did not join, but flowed independently into the developing D’Urville Canyon. During the Last Glacial Maximum, Fleming (1975) illustrated these rivers as combined on the emergent plain; similarly rivers from the South Island. The distance between running freshwaters would have been small, and not a great impediment to most adult simuliids. Further, given the flattened terrain, these rivers were probably large, braided, and with substrate of cobble and gravels, habitats quite suitable for at least current-day A. tillyardianum larvae, a species that occurs on both main islands.
At maximum sea level depression, Stewart Island would have extended southwest some 140 km, but does not appear to have incorporated Snares Island (Fig. 518). Both Auckland and Campbell Islands were many times their present size (ca 6 800 versus 626 km2, and ca 4 000 versus 113 km2, respectively) and each extended towards the northeast. There was, however, a marine gap between the southernmost extent of Stewart Island land and that of Auckland Islands of some 100 km. The distance between the then Auckland Island land mass and that of Campbell Island was some 200 km. In all, much smaller distances than at present.
Great Barrier Island off the northeast coast of the North Island was well connected to the North Island during sea level depressions — the Jellicoe Channel between it and the mainland is a mere 50–60 m in depth over much of its extent (not illustrated here). Therefore the present-day occurrence of A. australense on Great Barrier Island is to be expected.
To the north of the North Island the Three Kings Islands were separated by ca 11 km at maximum sea level depression (Fig. 516), so the occurrence of simuliids on these islands is not unexpected. However, only A. longicorne is present, and not the more ubiquitous A. australense. This is perhaps related to the ability of A. longicorne to survive in intermittent and slow-flowing streams, such as those found on the Three Kings Islands.
Historical Biogeography of New Zealand
New Zealand biogeography has long been of interest (e.g., Cockayne 1909; Kuschel 1975) and not just to New Zealanders. The country is considered important for the resolution of major questions regarding historical biogeography of islands; in particular, should New Zealand be considered as just an island, or as a continent (e.g., Gillespie & Roderick 2002; McDowell 2008; Wallis & Trewick 2009)? In recent decades, with the advent of increasingly sophisticated molecular techniques, research on New Zealand biogeography has undergone a major resurgence. The subject is extremely contentious and the literature considerable and oncoming. Recent reviews that provide an entry to the literature are by Trewick et al. (2007), Goldberg et al. (2008), Wallis & Trewick (2009), and Giribet & Boyer (2010). Broad topics dealt with revolve around what has been referred to as “Moa’s Ark” (Bellamy et al. 1990) where New Zealand as an ancient land mass possessed a similarly aged biota derived from its original formation as part of Zealandia, that is, a continental landmass which split off the eastern side of Gondwana some 120–80 Mya (Mid to Late Cretaceous). There is no question that proto-New Zealand, as part of Zealandia, possessed Gondwanan biota. Fossils point to this, an example being dinosaur fossils found on the Chatham Islands (Campbell & Hutchings 2007) and an array of dinosaur and plant fossils on the mainland (Molnar & Wiffen 1994, Worthy et al. 2011; Jordan et al. 2011).
However, geological evidence (Landis et al. 2008) that there was major peneplaining and marine erosion of the New Zealand landmass in the Oligocene (34–23 Mya) and a possible “Great Inundation”, has lead to re-examination of the age of New Zealand’s biota. That the country’s biota is largely ancient has now in part been discounted, and other scenarios suggested, such as New Zealand being the “Fly-paper of the Pacific” (Didham 2005; Giribet & Boyer 2010), “Goodbye Gondwana” (McGlone 2005), and “Hello New Zealand” (Trewick et al. 2007). These recent considerations have New Zealand being colonised by dispersal since the country rose from the sea after the “Great Inundation”. That is, the biota is much younger than previously considered, and has speciated rapidly under the influence of sequential glaciation (e.g., Plecoptera, McCulloch et al. 2010; and others).
As noted there are problems associated with organisms that do not appear to have any great dispersal ability, such as the iconic Sphenodon (tuatara), Onychophora (velvet worms), and the leiopelmatid frogs. Molecular examination of some of these and others (e.g., Allwood et al. 2010; Boyer & Giribet 2007, Buckley et al. 2011) indicates that they have a long history in New Zealand, suggesting that New Zealand was not entirely submerged during the Oligocene. This is in keeping with fossil evidence (Worthy et al. 2006, Worthy et al. 2011) from the southern South Island, which showed there was already a considerable vertebrate biota in the Early to Mid Miocene (19–16 Mya).
There are also problems associated with using single molecular markers for biogeographic purposes, undersampling which can produce spurious patterns of haplotypes, and unknown extinct lineages, which can affect inferred phylogenies. Of relevance here is that it is common to use but a single exemplar from a population to establish haplotype distribution, such as we have done in large part. Populations where more specimens were used show other haplotypes present. Consequently, our distributions of haplotypes and biogeographic comments must be taken as preliminary hypotheses.
Apart from the obvious barrier of modern Cook Strait to distribution of New Zealand biota, there are other well established broad distribution patterns that appear to indicate restriction of biota at times (e.g., Wardle 1963; Goldberg et al. 2008; Marra et al. 2009). The general term “beech gap” is sometimes applied to one such disjunction for the reason that in the South Island, the iconic Nothofagus (southern beech) is absent from a 300 km gap between north of Greymouth to south of Paringa (e.g., Trewick & Wallis 2001; Burrows 2005).
Trewick & Wallis (2001) used molecular data for 11 New Zealand invertebrates with gap distributions to test a priori hypotheses regarding these so-called “beech gaps”. There have been numerous other examinations of the Westland beech gap (e.g., Leschen et al. 2008) and the causes for other disjunctions (e.g., Marra et al. 2009; McCulloch et al. 2010).
From early times (Cockayne 1909) the general explanation for these gaps has been considered to involve Pleistocene climate extremes, namely glaciation and its effects. At considerable odds with that is a panbiogeographic explanation, involving geological movements along the Alpine Fault that forms the Southern Alps (Heads 1989, 1998). Both scenarios make different predictions regarding historical biogeographic relationships of taxa involved. In brief, if glaciation was involved, then sister taxa should be relatively young, if it were the Alpine Fault, then genetic distance between populations would be older.
Trewick & Wallis’ (2001) examination concluded that there was no evidence at all to support hypotheses involving the Alpine Fault and they proposed three broad patterns for the various distributions: Gap — north and south populations that have failed to unite; Colonisation — recently dispersed taxa that have closed the gap; Regional — distinct lineages that do not appear to reflect any history of a gap corresponding to Pleistocene glaciation. We add to these patterns; Refugia —restricted distributions that can be related to glaciation.
Similarly, Neiman & Lively (2004) and Neiman et al. (2005), as part of an in-depth examination of sexuality of the aquatic snail Potamopyrgus antipodarum (J. E. Gray), demonstrated that genetic divergence between lineages on either side of the Alpine Fault and Southern Alps were not involved in present-day distributions of the snail. Their data pointed directly to Pleistocene glaciation as the main factor for the current distributions. Their studies are rather unusual in that considerable numbers of snail individuals from each of the localities were subjected to DNA analysis (Neiman et al. 2005), rather than the normal single individual as is commonly done elsewhere and which we have done in large part for our study.
The biota of the New Zealand subantarctic islands has long been of interest, because of the marked endemism across and within the islands, and also the presence of apparent older flora and fauna. Of interest here, because of the presence of Austrosimulium, are the Auckland and Campbell Islands. Michaux & Leschen (2005) in a substantial “biogeological” analysis of the flora and fauna of the subantarctic islands, considered the biota to be a depauperate, unusual paleo-endemic biota mixed with recent precinctive forms. For biota they used as far as possible groups for which phylogenetic information was available. Their review of the paleogeology of the Campbell Plateau is, however, somewhat at variance with other previously discussed accounts (King et al. 1999; Cook et al. 1999). For example, while Michaux & Leschen acknowledged the presence of the Great South Basin, a feature which is considered by others to be the original rift of Zealandia from east Gondwana and has been open sea since that original rifting, they did not consider it to be a major barrier for biota between the Plateau and proto-New Zealand. After rifting from Gondwana the Campbell Plateau was above sea level and possessed a biota that has been well established by known coal seams (Zhu et al. 2006) and fossil pollen from Nothofagus and Araucaria (Michaux & Leschen 2005). As Zealandia rotated away from east Antarctica, the continental crust that constituted the Campbell Plateau was stretched and thinned, a warm mantle plume moved away towards the east, and the Plateau foundered. Timing for that seems well established, although the rate of submergence is not clear. Still, the process was rapid and by the Late Eocene (40 Mya) (Sutherland et al. 2010: their Fig. 2) the Plateau was under water. The problem for biota is that the volcanism which produced first the Auckland Islands and then Campbell Island, did not commence at earliest 19 Mya (Late Oligocene). There is little question that the islands were much more extensive than their current heavily eroded remnants. Further, they probably have submerged to some extent as is normal for cooling, aging, hot-spot islands (Craig 2003). Nevertheless, the question remains: where did the Plateau biota go for that considerable period of time?
In their discussion on biogeology, which involved interweaving biogeography, paleogeology, and paleoclimates, Michaux & Leschen (2005) invoked ideas from Gillespie & Roderick (2002) that dealt with the evolutionary consequences of reduction and fragmentation of landscape, so-called “relaxation”, to explain the overall depauperate nature of the biota of the Campbell Plateau. Michaux & Leschen were of the opinion that the Plateau sank at the Oligocene (ca 25 Mya) and that some land must have been above sea level until then to account for the persistence of paleo-endemics. No mention is made of the 16 million year gap between sinking of the Plateau and formation of the islands. Still, they illustrated (their Fig. 3) the juxtaposition between the Plateau and the southern portion of the South Island as it might have been at 71 Mya (end Late Cretaceous) and raised the point that perhaps the New Zealand alpine fauna might have originated from the cold-adapted Campbell Plateau paleofauna. However, they placed orogeny of the Southern Alps as commencing at 10 Mya; a time at considerable variance with the 5 Mya estimates of others (e.g., Chamberlain et al. 1999; Chamberlain & Poage 2000).
Michaux & Leschen (2005) in dealing with arthropods of these subantarctic islands included Plecoptera and Trichoptera. They deduced that the Plecoptera was a mixture of older endemic species with the possibility of other taxa being wind-borne from New Zealand with subsequent loss of wings. They also discussed Diptera, but not Simuliidae, even though there was a cladistic phylogeny available for Austrosimulium (Dumbleton 1973). Of interest is that loss of flight (brachyptery), common in subantarctic island Diptera, has never been observed for simuliids anywhere, even under extreme conditions (Craig et al. 2003).
Of relevance here for Austrosimulium is Michaux & Leschen’s suggestion that relationships of the fauna of the subantarctic islands is often with the southern South Island. We too see that relationship; however, given the molecular divergences in the ungulatum species-group, evidence from pollen rain (McGlone 2002), and tephra reaching subantarctic islands (Alloway et al. 2007), plus ages of those islands, we are firmly of the opinion that Austrosimulium reached Auckland and Campbell Islands by wind dispersal from New Zealand; in agreement with Dumbleton (1963b). That is, presence of the genus on these subantarctic islands is not related to a relictual Gondwanan fauna suggested for the Campbell Plateau.
Towns & Peters (1996) and Hitchings (2005) in discussing the leptophlebiid mayfly Cryophlebia aucklandensis (Peters), endemic to the Auckland Islands, noted that it had close affinities to New Zealand genera. Hitchings followed Michaux & Leschen (2005) in concluding that it was relictual and a remnant of a Campbell Plateau fauna.
Previous studies on Austrosimulium relationships
It has long been recognised (Edwards 1931) that there is a relationship between some members of Australasian and Chilean simuliids, originally Austrosimulium and Paraustrosimulium, and then Cnesiamima. This has been amply confirmed by morphological and molecular studies (e.g., Wygodzinsky & Coscarón 1962; Dumbleton 1973; Davies & Györkös 1988; Moulton 2003; Gil-Azevedo & Maria-Herzog 2007) and again here (Fig. 506). More recent re-examination (DAC, pers. obs.) of the enigmatic ?Austrosimulium colboi Davies & Györkös shows that entity is most likely to be the Australian representative of Paraustrosimulium. Possession of remarkable inflated pupal gills and well developed parameral plates in the adult male, both absent from Austrosimulium, point in that direction; a conclusion fully consistent with Moulton’s molecular findings. Such relationships are generally taken as indicative of vicariance mediated by the breakup of Gondwana in the Cretaceous. Further, a fully supported sister relationship (Fig. 506) between ?Paracnephia pilfreyi (Australia) and Cnesiamima atroparva (South America), strengthens that assertion for these Austral simuliids. Gil-Azevedo & Maia-Herzog (2007) also showed C. atroparva to be sister to Paraustrosimulium +Austrosimulium. Gibbs (2006) was of the opinion that New Zealand Austrosimulium was a “Ghost of Gondwana” and nothing appears to completely discredit that; however, as discussed below, it is more probable that New Zealand simuliids are “Ghosts of Gondwana” once removed.
Tonnoir (1925) commented briefly on geographic distribution of Austrosimulium, pointing out that with the state of knowledge at the time he could only be general. He noted that there were no species common to Australia and New Zealand. There was, however, a curious parallelism between Tasmanian Austrosimulium and New Zealand species, notably in female and pupal structures. Habitat choice of the larvae differed, however. He noted that A. australense and A. ungulatum were the most prevalent species in the North and South islands, respectively.
Dumbleton (1963b), in a major consideration of distribution of Simuliidae with emphasis on Austrosimulium, was not sure that New Zealand was colonised by Austrosimulium via aerial dispersal from Australia, given that it did not reach New Caledonia. He was of the opinion that presence of Austrosimulium’s sister taxon Simulium in Fiji and New Caledonia was the result of dispersal along the Melanesian Arc, and concluded that Austrosimulium reached New Zealand via overland connection, not significantly later than the Cretaceous.
Failure of Simulium to reach New Zealand was suggested as the result of the genus not reaching Australia from the north until after separation of New Zealand from Australia. That there are no species of Austrosimulium in common between the two countries, but one species-group that is, suggested to Dumbleton that New Zealand’s simuliid fauna arose from a derived Australian group rather than a more plesiomorphic group. Perhaps entry of Austrosimulium stock into Australia was from the north, their current absence from Malaysia and other parts of Asia paralleling the distribution of other taxa. He mused on why there are not more primitive simuliid taxa in New Zealand, and wondered if it was from filtering effects along dispersal routes. He also commented that Tonnoir (1925) was of a similar mind. One must remember, however, that in Dumbleton’s time (1963) events involving dismemberment of Gondwana were uncertain. Dumbleton did not specifically comment about absence of Australian Paracnephia from New Zealand, but lack of that as well as Simulium, could well be interpreted as indicating New Zealand Austrosimulium were derived from a disperal event.
Whereas Mackerras & Mackerras (1950) were of the opinion that Pleistocene cold climate was responsible for extinction of some biota, Dumbleton (1963) felt that Australian simuliids such as the furiosum species-group of Austrosimulium would not have gone extinct in New Zealand during cold periods since they are well adapted to such conditions, as are the sister taxa mirabile and ungulatum species-groups. He considered various scenarios for the southern bias in species numbers, noting the obvious barrier of Cook Strait, suggesting that the more active speciation in the South Island perhaps had been accentuated in the Pleistocene, but not necessarily confined to that period. Subsequent dispersal had not obscured the bias since there appeared to be little ecological reason why South Island species had not occupied the North Island. He also commented that the Auckland Islands species A. vexans must be derived from A. ungulatum via post-Pleistocene aerial colonists. Dumbleton observed that A. ungulatum appeared to need forest, which was largely absent from the Canterbury Plains and Banks Peninsula, and further that it was probably the oldest of the ungulatum species-group. As discussed later (p. 83), depauperate forest on the Peninsula is recent. At the time, Dumbleton felt that since A. australense had a markedly disjunctive distribution (then known only in the north and south of the South Island), it was probably a relict segment of a pre-Pleistocene species that had not managed to re-colonise the east coast of the South Island from either end; he suggested a similar scenario for A. laticorne.
Later, Dumbleton (1970), in a broadly-based Presidential address given to the Entomological Society of New Zealand 1969 conference, considered Pleistocene climates and their effect on New Zealand insect distributions. He mentioned a problem in dealing with taxa of warm-temperature adaptations (Malayo-Pacific, Australian) and those which were of apparent southern origin and cold-temperature adapted. He noted that it would be very bold to assume that species that appeared to have remained morphologically consistent through the Tertiary were unchanged in temperature tolerance. Another problem of uncertainty that he noted was the rate of speciation and the morphological differentiation by which species were then recognised. Specifically referring to Austrosimulium and the Australasian–South American relationship, he noted that the rate of divergence since the Cretaceous had been remarkably slow. He did not elaborate on this point; however, it appears he was of the opinion that Austrosimulium was of Cretaceous origin.
The South Island was commented upon as having more species of biota than the North Island and Dumbleton suggested that being 30% larger in area, and greater range of habitats, might provide the explanation. He considered the possibility of extinctions caused by extreme fluctuations in temperature during glaciations, noting that at that time there was no evidence for this for insects, but that plants seemed to show it. Shifts in distribution centred mainly around the effect of Cook Strait, and he noted a number of examples where the Strait formed the southern boundary to distribution. Physiological adaptations were admittedly puzzling at that time, in particular those involved with the alpine biota of the South Island. The main question revolved around how did the biota survive warm periods if the organisms were ancient, and the Southern Alps not? If the biota was of more recent origin and colonised colder habitats when they became available, from where did the organisms originate? Dumbleton also raised the question of how wingless forms of insects would arrive in New Zealand.
Some suggestions were made regarding the influence of refugial area, particularly in the South Island. He then dealt with distribution of a number of insects, including Austrosimulium, providing a map (1970, his Fig. 7) of the then known distributions of the various species of simuliid. His distributions of Austrosimulium spp. are little different from those presently known. Of note, however, is that A. tillyardianum is now known from much farther north in the North Island (Map 14) and A. australense is more widespread (Map 3) in the South Island.
In his seminal work on Austrosimulium, Dumbleton (1973) re-examined various aspects of origin and distribution of Austrosimulium. He judged Austrosimulium and its sister genus Simulium to be of equal evolutionary age and derived from lower simuliids. He considered the origin of Austrosimulium and whether it was derived from the same ancestral stock as Simulium, or had had an earlier origin perhaps from Cnephia or Gigantodax, and if similarities to Simulium were due to convergence. The anal sclerite of the larva was singled out for considerable examination. One possible origin he considered was that Austrosimulium arose in Antarctica, and was derived from Gigantodax. However, he reiterated that New Zealand Austrosimulium had their main relationship with Australia. Consideration was also given to Brundin’s (1966) suggested relationships for Podonominae midges. Dumbleton showed that Austrosimulium did not follow Brundin’s rules, in particular that stating: “there are no direct phylogenetic connections between a group of Tasmania-Australia and a group of New Zealand”. Dumbleton suggested one scenario where the origin was an ancient trans-Antarctic dispersal from Australia to South America, in which case Paraustrosimulium and Novaustrosimulium would be plesiomorphic segregates and Gigantodax-like features (i.e., the semicircular sclerite of the anal sclerite) would be independent developments.
In summarising his ideas on relationships for his Fig. 252, Dumbleton concluded that Paraustrosimulium (South America) and Novaustrosimulium (Australia) were products of ancient geographic isolation. The close morphological relationship between Australia and New Zealand Austrosimulium suggested that the two countries shared undifferentiated stock of that subgenus. The ungulatum species-group shared by both countries could be attributed to the isolation of New Zealand with parallel evolution subsequent to that. He further concluded that the three segregates of Austrosimulium (as he recognised them) had been isolated since the Cretaceous. He noted for Australia that aridity would have imposed a rigour on evolution, whereas in New Zealand glacial climates would have been a driving factor. The unicorne-subgroup probably arose though physiological adaptation to greater cold. Geographical isolation was also no doubt involved in the restricted distribution of A. (A.) fulvicorne (Frazer Island, Australia), A. vexans (Auckland Islands), and A stewartense (Stewart Island), and the species were probably of Pleistocene age. He made similar observations for A. fiordense, then considered to be a subspecies of A. multicorne. He noted that the relationship between species of the tillyardianum-subgroup would need further investigation, probably cytological – a prescient observation given our results from molecular analysis of this subgroup.
McLellan (1975) in an overview of the freshwater insect fauna of New Zealand noted that in general relationships were firstly with Australian fauna and secondarily with South American. He was, however, clear that some chironomids had connections to South America in the first instance. Fleming (1975) also emphasised the relationship of New Zealand biota to Australia firstly, and noted clear evidence for dispersal as the origin of some biota.
Cranston (2005) while dealing with historical biogeography of Diptera noted that the Parochlus chironomid group has, firstly, strong New Zealand + South American connections, and secondarily, connections with Australia. He mentioned that Austrosimulium fitted what is referred to as the “Trans-Tasman track” by which dipterous fauna shows links between eastern Australia (including Tasmania), New Zealand, and New Caledonia. The overall distribution of Austrosimulium fits that track well, although the genus is absent from New Caledonia, and occurs on subantarctic islands. Cranston invoked an ancient geological scenario as an explanation for the Trans-Tasman track, namely Cretaceous timing for the disconnection of New Zealand from Antarctica. He noted that this was surely associated with speciation patterns seen in several New Zealand taxa. Further, other biotic links to Australia may indicate a hybrid origin of biota via dispersal, extinction, or ecological responses as the respective land masses moved northward.
Recently, Cranston et al. (2010) used multiple genes and sophisticated analyses to examine relationships of podonomine chironomids in the southern hemisphere. They were quite definite that all their sampled New Zealand taxa supported a South America + New Zealand relationship, and not one with Australia. Their dates of divergence (83-53 Mya) of some lineages supported vicariance (Cretaceous break-up of Gondwana) as the basis for distribution, and indicated that some New Zealand taxa must have survived the Oligocene Inundation. Even more recently, Krosch et al. (2011) used similar techniques to examine Gondwanan Orthocladiinae midges. They found a complex situation with some lineages showing the expected vicariance distribution from break-up of the super continent, and, as above, evidence that lineages survived inundation (if any) during the Oligocene. Further there were at least 3 examples of trans-Tasman Sea dispersal events post-dating the Gondwanan break-up.
Our molecular examination of Austrosimulium was restricted to New Zealand and does not show deep divergences, even of disparate lineages (Fig. 509a, 509b) that would indicate a Cretaceous origin. However, as mentioned elsewhere, work on Novaustrosimulium involving the 12S RNA gene (Ballard 1994) indicated that the Australian segregate is older. Our morphological cladistic analysis (Fig. 505–507) is not at variance with that, and adds little to what has already been suggested for Austrosimulium, i.e., the genus is sister to Australian + South American taxa. Still, relationship between the subgenera Novaustrosimulium and Austrosimulium is moot. In Strict Consensus the former did not resolve as a single clade (paraphyletic), indicating that taxonomy of that subgenus needs revision. It should be noted that Dumbleton (1973) erected Novaustrosimulium as a necessity to avoid amending the diagnosis of Paraustrosimulium within Austrosimulium.
A difference we see in our morphological analysis from that done by Dumbleton (1973) is that where he considered the Australian mirabile species-group as sister to the Australian plus the New Zealand ungulatum species-group (his Fig. 252), we show the Australian mirabile plus ungulatum species-groups to be sister lineages, and together sister to the New Zealand ungulatum species-group.
We note basic limitations of our phylogenetic investigations. The morphological analysis lacked data for many of the Australian species and the molecular analysis had no Australian data. Similarly, various New Zealand species were not included. For the molecular analysis (Craig & Cywinska 2012, pp. 60–65), material was not available for A. campbellense, A. extendorum, or A. fiordense, and the CO1 gene did not provide useful discrimination between species of the tillyardianum-subgroup (see Fig 508a, 508b). Austrosimulium vailavoense was not included in morphological considerations because the immature stages and adult males are not known. It is clear that Australian Austrosimulium needs to be taxonomically revised, with descriptions brought up-to-date, and analyses repeated with the full slate of species. Of importance would be molecular analysis.
The only molecular analyses of Austrosimulium species are those of Ballard (1994) and Moulton (1997, 2000, 2003). Ballard used the ribosomal RNA 12S gene which was congruent with morphological differences between Austrosimulium (N.) pestilens and A. (N.) bancrofti. However, he could not resolve cytological forms known for A. bancrofti (Ballard & Bedo 1991).
Current molecular study on New Zealand Austrosimulium
Analysis of the mt16S RNA gene by Craig & Cywinska produced poor discrimination between species of New Zealand Austrosimulium, was of little utility, and is not shown here (however, it is available in Supplementary Material on the website <fnz.landcareresearch.co.nz>). Given the generally assumed conservative nature of ribosomal genes (Simons et al. 1994; Trewick & Wallis 2001), Ballard & Bedo’s (1991) results indicate that Novaustrosimulium is older than New Zealand Austrosimulium, in agreement with its placement in the morphological cladistic analysis (Fig. 505, 506). There is, however, no inherent information in the cladistic analysis regarding timing of arrival of New Zealand Simuliidae. That is, is it vicariant and as old as New Zealand’s geology, or is it younger and an arrival by dispersal after the Oligocene?
The molecular analysis of the CO1 gene (Fig. 508–514) for 15 of the 19 recognised species of New Zealand Austrosimulium by Craig & Cywinska (2012, this monograph) supports the latter, a dispersal scenario. Concatenation of the large tillyardianum species-group (Fig. 508a, 508b) strongly indicates that it is a segregate of recent origin, and is in strong agreement with the poorer resolution obtained for this assemblage in the morphological cladistic analysis (Fig. 505–507).
Furthermore, most species of the tillyardianum species-group do not occur in the North Island, suggesting that the group probably arose in the South Island.
Ages of CO1 haplotype lineages are often calculated from the rate of molecular change of the gene over time. A rate commonly applied for animals is 2–3% per million years (McCulloch et al. 2010). For arthropods, 2.2–2.3% per million years has been suggested as the rate to use (Brower 1994; Gaunt & Miles 2002). Trewick & Wallis (2001) used 2.3% for their biogeographic analyses, and we do here too. There are problems with this approach. In particular, rates of change may not be the same in different lineages (Avise 2004a), and the rate of change may not be constant in any one lineage. For Austrosimulium, one problem relates to individual species having different numbers of generations per year, from defined univoltine to multiple generations, some unsynchronised, even to a mixture depending on location. Genetic divergence accumulates generation by generation, so voltinism of species is of some importance. Application then of a single rate of genetic divergence to lineages, some strictly univoltine and others with unsynchronised multivoltinism, will result in mis-estimations of time of divergence of lineages.
We follow a commonly used practice for New Zealand biogeography, which is to determine divergence between lineages and then apply the estimated age to the geographic/geologic history of the islands (e.g., Trewick & Wallis 2001). More recent works (e.g., Cranston 2005; Smith et al. 2006) are most clear that each species should be biogeographically considered separately, in part because dispersal abilities of taxa differ. When similar patterns between disparate organisms emerge, a common deterministic event may be invoked, for example, glaciations, sea gaps, volcanism (e.g., Trewick & Wallis 2001; Goldberg et al. 2008).
The backbone of the rooted Neighbour Joining (NJ) tree from the molecular analysis of New Zealand Austrosimulium (Craig & Cywinska 2012, this monograph) is largely consistent with that of the morphological character analysis (Fig. 506, 508a, 508b, 509a, 509b), probably for the very reasons that Cranston et al. (2010) obtained similar concordance for chironomids — namely, use of characters from all stages. As noted by Craig & Cywinska (2012, this monograph), the major difference in topology between our morphological and molecular analyses is that Dumbleton’s unicorne-subgroup (ungulatum species-group) is not sister to the ungulatum-subgroup, rather to all of the remaining New Zealand Austrosimulium. Further, in the tillyardianum-subgroup NJ analysis does not usefully group morphologically distinct species, such A. tillyardianum, A. laticorne, and A. multicorne (Fig. 508a, 508b); similarly, the cladistic Strict Consensus tree (Fig. 505) that also resolves the tillyardianum-subgroup poorly. Dumbleton (1973) too was of the opinion that this group was in need of special attention. Lack of resolution in the molecular analysis of the tillyardianum-subgroup tends to discredit clustering of haplotypes such as occurs with A. tillyardianum (Fig. 508a, 508b). We comment on this later.
Historical Biogeography of New Zealand Austrosimulium species
australense species-group
australense-subgroup
Austrosimulium australense
The overall distribution of A. australense (Map 3) shows no obvious gaps in the North Island. Indeed, the species appears relatively evenly distributed, indicative of good dispersal. However, that assertion contradicts a previous one, namely adults of this species are not found far away from breeding localities. The South Island segregate shows a distinct gap from the mid Canterbury Plains, through Otago to southern Southland. The southern distribution is markedly similar to that of other New Zealand taxa (Trewick & Wallis 2001; McCulloch et al. 2010) and fits the definition of a gap distribution. Absence from Fiordland may be real, but is as likely to be from lack of collecting. Dumbleton (1973) was of the opinion that overall distribution of A. australense was the result of recent glaciation with subsequent lack of dispersal by either of the northern or southern segregates of the species to fill that gap. This is still the common explanation for such patterns in South Island biota (e.g., Trewick & Wallis 2001). However, there is no gap for A. australense in Westland, so dispersal since the Last Glacial Maximum appears to have occurred there for this species. Perhaps ecological factors are also involved, since this species is typically found in open streams with trailing vegetation. Where there is a lack of vegetation A. australense is generally absent. Dumbleton (his p. 546) also commented about ecological factors, such as humidity and vegetation, determining distribution.
In the molecular analysis, haplotype #1 (NZS103) (Fig. 510) is sister to the australense species-group, but was morphologically confirmed as A. longicorne! Still, that collection was of scant material of immature larvae and difficult to identify, so may be a misidentification. Divergent at some 3.4% it may well represent an undescribed species. However, the sister relationship and occurrence on Takaka Hill, northwest Nelson, might be of significance. Apart from the area being geologically ancient (Cretaceous limestone) it is generally thought not to have been extensively glaciated, including the Last Glacial Maximum (e.g., Alloway et al. 2007). That region, in general, also had the most haplotypes (#s 2, 3, 4, 5, 6) of A. australense. Such diversity of haplotypes in an area is often taken as an indication of longer evolutionary time — conversely, populations with low divergence of haplotypes are considered to indicate recent dispersal. Does this diversity indicate A. australense and its various haplotypes perhaps arose in the northwest quadrant of the South Island?
Genetic divergence between the North and South Island segregates of A. australense was some 6%, giving an age for separation of 2.6 Mya (end of the Pliocene). At that time in New Zealand the Manawatu Strait was still in existence, but becoming shallowed and with remnants of the Wanganui–Marlborough Shield present. By 2.0 Mya the Strait was narrowing with mountain ranges occluding it to the east (Trewick & Bland 2011). This scenario suggests A. australense moved northwards along those new mountains when the Manawatu Strait eventually was closed, diverging later to become the North Island clade.
Two populations of the southern A. australense (NZS170, Kaipipi Inlet, Stewart Island, Fig. 511; NZS 29, Catlins, South Island, Fig. 512) for which molecular data are available, are related to separate haplotypes from the North Island, not the South Island. Is this indicative of an even older disrupted distribution or, errors in analysis? Assuming the former, those populations are divergent from North Island A. australense at about 1.6%, which would equate to some 700 000 years BP (Mid Pleistocene). At that time the North and South Islands were connected across the modern Cook Strait at the Marlborough and Wellington regions and the Manawatu Strait was well gone (Lewis et al. 1994, Trewick & Bland 2011). An assumption here is that subsequent glaciations and concomitant ecological changes produced that eastern gap to the south. The Stewart Island segregate, and that from the Catlins, of A. australense, have habitats for immature stages that are at considerable variance to those of more northern exemplars — namely heavily shaded streams. The southern segregate of A. australense, if a relictual population, deserves further investigation, perhaps with finer examination of morphological structures, as was done for cryptic species of simuliids in Britain (Day et al. 2008).
With few exceptions, North Island haplotypes of A. australense (Fig. 511, 512, 513) have a similar distribution, namely, concentrated in Northland and Coromandel, east coast and sporadically in the central regions. Again, this is a well established pattern for New Zealand biota. However, A. australense overall occurs in all 13 aquatic ecoregions of the North Island.
The concordance between distributions of North Island A. australense haplotypes (Fig. 511, 512) and those for other insects, for example, the tree weta Hemideina thoracica (White) (Goldberg et al. 2008: their fig 6) and the caddisfly Hydropsyche fimbriata (McLachlan) (Smith et al. 2006) is strong. Such similar patterns shared by such widely diverse organisms are evidence of a common strong evolutionary event.
Goldberg et al. (2008) suggested that the distribution for the weta is consistent with Pliocene islands being present to the north at that time, with haplotype divergence occurring on the islands, in particular with the Manukau Strait across the present Auckland region acting as a barrier. Smith et al. (2006) also invoked this scenario, in part, for distribution of H. fimbriata haplotypes, suggesting a mid Pleistocene (780 000 years BP) divergence. They also considered impact of volcanic eruptions on aquatic organisms.
Volcanic ash and pyroclastic flows can be devastating for stream-dwelling organisms as is well known for the Mount Saint Helens eruption, on the west coast North American, in 1980 (Anderson 1992). Recovery from total removal of aquatics by scouring was rapid for the St Helens streams. In New Zealand, a light volcanic ash fall had little effect on the Tongariro River after the Ruapehu eruption of 1995 (Collier 2002). McDowall (1996) examined in detail the effects of a major eruption from Lake Taupo, some 2 200 years BP, on fish fauna of the northeastern part of the North Island. He was of the opinion that rivers which were impacted would likely still reflect that disruption to their fauna, not just to fish but also to the invertebrates. He noted that there was little indication of aquatic invertebrate endemicity in the volcanic region which might indicate antiquity and survival through such eruptions. For simuliids, however, there appears to be recent endemicity of A. dugdalei. We also wonder if the dense and widespread populations of A. longicorne on the volcanic plateau, where the larvae are found in specialised habitats associated with seeps high on the slopes of the volcanoes, may be an instance of survival in refugia and then good dispersal skills. We have alluded to the good dispersal skills of A. longicorne elsewhere (p. 113). Further, is it possible that the abrupt change in heterozygote inversion pairs on A. australense chromosomes on either side of the Tarawera River (McLea & Lambert 1983, 1985) is a result of that volcanic disruption?
Later, McDowall (2005) suggested that the crayfish Paranephros would have been extirpated by volcanic activity, but recolonised rapidly. Be that as it may, a series of such events in the North Island must have well sterilised the central regions. These were eruptions from a now-filled caldera north of Lake Taupo. The largest was at ca 1 Mya and produced the (Cape) Kidnappers Ignimbrite, a term used for the distinctive remains of pyroclastic flows, and was one of the most extensive pyroclastic flows ever known for Earth. Other debris from that eruption is referred to generally as the Potaka tephra (Wilson et al. 1995; Wilson & Leonard 2008).
Wilson et al. (1995) noted that the Taupo Volcanic Zone has produced some 34 ignimbrite eruptions from some 8 calderas over the last 1.6 my. The pyroclastic flow at 1 Mya reached as far north as Auckland and well south onto the Wanganui Basin. Its extent into the northern Hawkes Bay region and that of East Cape is uncertain, but deep sea drill cores indicate that, in places, the flow went well out to sea. Higher areas, such as the Bombay Hills and Hunua Range south of Auckland, Coromandel and Kaimai Ranges to the east, Hapuakohe Range, Taupiri and Hakarimata Ranges, and farther south were spared, as were peaks just northeast of Hamilton.
The correspondence between that pyroclastic flow and distribution of A. australense haplotypes # 12 and # 19 is marked, and similarly, but less so for that of #17 and #18. So, we suggest that the present distribution of those haplotypes is of upper Lower Pleistocene age (ca 1 Mya) and resultant from volcanic activity. This does not discount the possibility of involvement of the Manukau Strait as a dispersal barrier.
However, not only were there pyroclastic flows, but volcanic ash (tephra) was common. The recent Kawakawa (Oruanui), Okareka, and Rerewhakaaitu Tephras (ca 26 500, 19 100, 17 600 years BP, respectively) covered large areas and are useful in dating strata (Shane 2000). The first of these eruptions laid down 20 mm of ash (up to 2 000 mm in places) from south of Auckland, all over the North Island, northeast on the South Island, out to the Chatham Islands and even some to the Auckland Islands (Manville & Wilson 2004; Lowe et al. 2008; Manville et al. 2009). The thickness of some of these deposits is hard to imagine unless actually seen.
The centre of the North Island has been constantly sterilised, disrupted, and disturbed since the commencement of volcanism in the Taupo Volcanic Zone and the Central Volcanoes (perhaps 3 Mya and 0.27 Mya, respectively). The disruption caused by pyroclastic and tephritic activities is well reviewed by Manville & Wilson (2004) and Manville et al. (2009) and is suggested reading. Initial perturbation to fluvial systems (running water) is extreme, but so are the long-term effects on vegetation. Of considerable importance is that these effects depend on climatic conditions. For example, effects of the Oruanui eruption, which covered most of the North Island at 26 500 years BP during, the Last Glacial maximum, were still in progress ca 17 000 years BP and causing major reorganisation of running water systems and geomorphology. Is it possible, then, that the basic pattern of distribution of most of the A. australense haplotypes, that is, the generally low diversity in the central region of the North Island, is a result of the recent tumultuous history, rather than earlier events? Are we mis-estimating times of divergences between haplotypes? Austrosimulium australense has more than two generations a year, and in places they are poorly synchronised (see p. 107, species bionomics).
Haplotype #20 (NZS99, Green Hills) from the Farewell Spit region, South Island is of significance. This is sister to haplotypes #17–19 that occur up the east coast and in Northland. Given that haplotype #20 is based on more than one specimen, we are assured that this is not an error, and note that another North Island haplotype also occurs nearby at Totaranui (#31, NZS102, Fig 513). However, that their North Island sister haplotypes are more northerly raises a question — were intermediate populations removed by volcanic activity? We suggest a high probability for that.
Divergence for the Green Hills population is in the order of 550 000 years BP and that for the Totaranui population 325 000 years BP, indicating that there were two separate dispersal events from separate North Island stocks. Both suggested events could have occurred at times when the two islands were connected during glaciations (e.g., Lewis et al. 1994). For the Green Hills population, time of divergence from its North Island sister haplotype indicates the Kawhaka Glaciation, and for the Totaranui population the end of the Nemona Glaciation. Further sampling in that region and DNA analysis might be revealing.
Of significance is that these two populations have mixed haplotypes (see also Fig. 510); Totaranui also has South Island haplotype #5 and Green Hills has #2. Does lack of mixing of the two haplotypes indicate, as we have alluded to elsewhere, that the South Island segregate of A. australense is actually a cryptic species? A further question, too, is why have no South Island haplotypes of A. australense managed to reach the North Island?
Another haplotype with broad North Island distribution, but distinctly peripheral (Fig. 512), is #23 that consists of a markedly homogeneous series of populations. Indeed, it perhaps should be considered as just part of a single haplotype comprising numbers 23, 25, and 26. With the exception of one population of haplotype #26 (NZN92, Ohakune) and that of #22 (NZN30, Mangaweka, Rangitikei River), all are peripheral. Again, was volcanism from the Taupo Volcanic Zone responsible?
Slightly divergent from the above haplotypes is #24. This is known only from two populations in the south of the North Island and has an apparent relationship with a population (NZS29) from the Catlins, southern South Island! As for the relationship of the Stewart Island A. australense haplotype #14 to sister haplotypes in the North Island (Fig. 511), this is something of a conundrum and needs further examination.
Along those lines, is it possible that the widespread, homogenous haplotype #27 (Fig. 513) has colonised areas since the last major extirpation by volcanic activity, for example that by the Kawakawa (Oruanui) eruption (ca 26 500 years BP)? Effects of that eruption were still acting 17 000 years BP, as mentioned, previously (Manville & Wilson 2004).
McLea & Lambert (1983) examined the cytogenetics of A. australense and suggested three zones of inversion polymorphisms (see p. 108). They noted a marked cytological change between populations north of the Tarawera River and elsewhere. There also appeared to be differences in habitats occupied by immatures of the cytotypes. We have seen no such patterns in distribution of haplotypes, or habitat differences, but our samples around the Tarawera region were not extensive. Volcanics might also be implicated here. Manville & Wilson (2004) discussed major changes in watersheds for rivers flowing out of Lake Taupo following the Oruanui eruption. Similarly, the Tarawera River had massive break-out flooding and scouring during an eruption in AD 1315, and again in 1905, following the eruption of Mount Tarawera in 1886 (Hodgson & Nain 2005).
Austrosimulium longicorne
Currently known localities (Map 11) for A. longicorne show a scattered distribution with apparent gaps in both main islands, absence from Stewart Island, and presence on the Three Kings Islands. It is not obvious if these are strict gap distributions as such since the ecological requirements for the larvae of this species are distinctive: namely, smoothly flowing, constant velocity water with trailing vegetation and, as discovered recently, thin films of water with markedly low velocity in seepages. Further, populations of the species can survive in intermittent streams and, as we have suggested elsewhere, the dispersal ability of the species must be considerable. That is probably the reason that it, rather than the ubiquitous generalist A. australense, managed to colonise the Three Kings Islands. That colonisation probably occurred during a cool period when, for example, the distance between the extended North Island and the Three Kings was as little as 11 km during the Last Glacial Maximum (Fig. 516).
Molecular evidence shows (Fig. 510) that the species arose from the South Island segregate of A. australense. The divergence is ca 2.6%, and indicates an age of origin of ca 1 Mya (Mid Pleistocene). Morphological divergence, particularly between the pupal gills of the two species, is marked (cf Fig. 268, 269) and, as referred to elsewhere, is a good example of a lack of concordance between divergence rates of morphology and molecular aspects of species (p. 63 Molecular Results and Discussion). Haplotype #8 (NZS14, NZS41) has a marked altitudinal range (Christchurch and Old Man Range), that is perhaps suspect. Neither locality can be considered old, both being only a few thousands of years, if that. Haplotypes #9 and 9a (the latter moderately divergent) are arrayed only along the east coast of the South Island, but this is probably an artifact since the analysis lacks material from known populations (Map 11). The sister haplotype, #10, is only in the North Island, and is homogeneous and widespread. Divergent from haplotype #9 at some 1.5%, this indicates the split was at ca 0.87 Mya (lower Mid Pleistocene), when the connection across Cook Strait was narrowing and land ranged up to the Manukau Strait in the north (Lewis et al.1994, Trewick & Bland 2011). That there is but that single haplotype in the North Island is indicative that this dispersal pattern is recent.
tillyardianum-subgroup
Molecular evidence for this segregate concatenates species, and therefore has weak credence. Morphological divergence in some species is strong however.
Austrosimulium tillyardianum
The distribution of A. tillyardianum in the South Island (Map 14) is similar to that of A. multicorne and shows no gaps that relate obviously to glaciation. The apparent absence of A. tillyardianum from mid and south Westland is more likely the result of ecological factors. This species is typically found in open, clear-water streams with cobble substrate — a habitat not generally found in Westland where large rivers tend to have turbid water, high velocity, and unstable bed substrate not suitable for A. tillyardianum larvae. We too are of the opinion that its absence from southern South Island is probably related to temperature. The majority of known localities are in 10–12°C mean temperature areas (NIWA 2008) and this perhaps reflects in the species’ absence from Southland Plains and High Country aquatic ecoregions where the mean temperature is some 2 degrees lower. A dearth of A. tillyardianum localities from the central Marlborough region is likely to be an artifact of poor collecting, as the species is now known from the Awatere and Waihopai Valleys (NZS182–184).
Our expectation was that Banks Peninsula populations of A. tillyardianum might show, at minimum, subspecific differences to those in the remainder of the South Island. Other aquatic invertebrates have precinctive species on the Peninsula, and Winterbourn (2008) listed 5 species of caddisfly, a stonefly, a mayfly, and a blepharicerid, Neocurupira chiltoni. Craig (1969) was of the opinion that this blepharicerid was derived from N. tonnoiri Dumbleton, wind-blown from the west. Similarly, Hitchings (2008), when discussing post-glacial distributions of New Zealand Ephemeroptera considered Nesameletus vulcanus Hitchings & Staniczek endemic to Banks Peninsula, to be related to N. austrinus Hitchings & Staniczek which is found to the west in the Southern Alps. McLellan (1975) also commented on this. A precinctive biota agrees with isolation of Banks Peninsula as a volcanic island until some 75 000–20 000 years BP, when gravel outwash from glaciers filled the sea gap (Shulmeister et al. 1999; McCulloch et al. 2009). Dates given for the junction of the Banks Volcanoes to the mainland to form the peninsula are various. For A. tillyardianum though there are no obvious morphological characters that indicate any divergence due to isolation. This is in full agreement with the lack of molecular divergence within the tillyardianum species-group (Fig. 508a, 508b), again suggestive of recent origin.
In the North Island, A. tillyardianum is widespread from the Auckland region southwards. Our collections have extended, by a considerable extent, the range known to Dumbleton (Map 14). It is unlikely this represents recent expansion of the species’ range, but rather is the result of more extensive collecting. We make that assertion since it appears the distribution of A. tillyardianum is constrained by temperature, its northern extent falling generally in the 12–14°C mean annual temperature range (NIWA 2008). In the North Island, one gap in distribution that is probably real, even though it also represents patchy collecting, is that encompassing the Taranaki region. Streams and rivers there are, in part, cut down through blue mudstone (papa) and hard cobble substrates — typical habitat for A. tillyardianum larvae are generally lacking. The same applies to the Wairarapa. Similarly, A. tillyardianum habitats are sparse in the Waikato, however, for different geological reasons and because of heavy agricultural usage.
Lack of divergence (Fig. 508a, 508b) in the CO1 gene means we cannot usefully comment on A. tillyardianum distributions from a molecular viewpoint. Such homogeneity can indicate recent origin and rapid dispersal. The presence in the central part of the North Island, a region repeatedly disrupted by volcanism, also indicates considerable dispersal ability.
Austrosimulium alveolatum
Originally considered a subspecies of A. laticorne, but herein raised to species status, A. alveolatum occurs in a restricted area around Porters Pass in Canterbury (Map 2). Most likely derived from the A. laticorne western population (Map 10), A. alveolatum could be considered a break-out population through the passes in the Southern Alps. Such areas have been ice free for a relatively short time, possibly much less than 13 000 years BP (Burrows 2005).
Both A. albovelatum and A. alveolatum show distributions that would be expected from isolation caused by glaciers, as evinced for stonefly (Plecoptera) vicariant distributions by McCulloch et al. (2010). There is evidence that at maximum advance the Rakaia Glacier (Burrows 2005; Alloway et al. 2007) extended as far east as present-day Methven. Similarly, the Rangitata Glacier to the south extended to near Mayfield. Thence both cut through ecological zones on the extended coastal plains. Burrows (2005) and others (e.g., Soons 1994) showed that even at the final advance the eastern glaciers (Waimakariri, Rakaia, and Rangitata) terminated just upstream of their present-day gorges. During the Waimaunga Glaciation (ending 220 000 years BP) the Waimakariri, Rakaia, and Rangitata glaciers extended even farther eastwards than during the Otira Glaciation (Burrows 2005). Isolation caused by glaciations has often been invoked to explain distributions of biota (e.g., Hewitt 1996; McCulloch et al. 2010): we are of the firm opinion that this involves A. albovelatum and A. alveolatum, and that their current distributions are refugial. Such restricted areas are analogous to the “kipuka” of Hawai’i — regions of older land surrounded by fresh lava where speciation has taken place. Austrosimulium albovelatum and A. alveolatum are not closely related phylogenetically, so the glaciers acted on separate precursors.
Austrosimulium multicorne
This species occurs sporadically along the western side of the South Island and there appear to be no gaps as such (Map 12), except perhaps southern Westland. On the eastern side of the Southern Alps, the species is common along the foothills west of the Canterbury Plains and shows no evidence of disjunct distribution. Aggregation of localities hard along the Southern Alps is a collection artifact resulting from proximity of running water to roads. With lack of molecular discrimination within the tillyardianum-subgroup we cannot make useful comment from that point of view, except again, that lack of internal divergence indicates recent origin of the subgroup. Superficially though, the distribution (Map 12) of A. multicorne appears to fit well the “colonisation” distribution of Trewick & Wallis (2001), indicative of recent events since the Last Glacial Maximum and in full agreement with the molecular data (Fig. 508a, 508b).
The low frequency occurrence of A. multicorne on Banks Peninsula is of interest since the species’ presence there was unknown until now. Crosby (1974a), during his intensive work there on A. tillyardianum, did not recover A. multicorne. Neither did Dumbleton. It is unlikely that its presence is the result of recent colonisation, although that is possible. Its low frequency may reflect competitive exclusion by A. tillyardianum, which is widespread and occurs in astronomical numbers at some localities.
Austrosimulium dugdalei
Originally considered to be the North Island representative of A. multicorne by both Tonnoir (1925) and Dumbleton (1973), divergence of A. dugdalei from A. multicorne can perhaps be suggested as no older than 275 000 years BP, given that its precinctive distribution seems to rely on the high altitude of the Volcanic Plateau and southern Taupo Volcanic Zone (Map 6). The Volcanic Plateau consists of debris originally from the Tongariro volcano, and later from Ngauruhoe, Ruapehu, and other eruptions. Given that A. dugdalei appears to prefer higher altitudes, divergence from A. multicorne is no doubt more recent than given above, since time was required for those volcanoes to produce the plateau. Although data are limited, A. dugdalei does not group with A. multicorne in the molecular analysis (Fig. 508a, 508b), and suggests a separate species. Again, with little credence, the small amount of genetic divergence for A. dugdalei indicates an origin ca 250 000 years BP, congruent with the age of the Volcanic Plateau.
Austrosimulium albovelatum
Austrosimulium albovelatum has a similar distribution to that of A. alveolatum, being restricted to the eastern side of the central Southern Alps, from Porters Pass, south to the Rangitata River (Map 1). There is, however, an unusual northern outlier (NZS11) at Kaikoura, based on a pupa and cocoon, found at sea level (p. 116). Are there populations between Kaikoura and Porters Pass? All populations fall into the High Country aquatic ecoregion, a finger of which extends out to Kaikoura from the mountain range, so we assume there will be.
Apart from this Kaikoura outlier, the body of the distribution indicates an origin in a refugium between major glaciers (Waimakariri and Rakaia), and not necessarily from the recent Otira Glaciation. The arrangement of glaciers (Burrows 2005, his Fig. 70) for the Waimaunga glaciation (300 000 years BP) was similar to that of the Otira.
Austrosimulium laticorne
Although widespread in the South Island, A. laticorne is precinctive and is unknown from North and Stewart Islands. It is widely distributed on the West Coast, northwest Nelson and into the Marlborough Sounds area, but is mostly absent from the rest of the island except for the most southern part (Map 10), indicating a gap distribution rather similar to the South Island A. australense (Map 3).
The northern population might indicate the Southern Alps acted as a barrier and that the population has not yet managed to disperse east. That central Marlborough region is poorly collected, and the current known distribution may well be an artifact as suspected for other species. Recent collections in the Marlborough region failed to discover A. laticorne, but they extended the distribution for A. tillyardianum. The outlier population east of the Southern Alps at Lake Tekapo may well represent penetration of the Southern Alps by the Westland population. Such penetrations have been suggested for the cicada Kikihia subalpina (Hudson) (Marshall et al. 2009), and elsewhere across the Southern Alps for galaxiid fish as river capture alters drainage basins (Craw et al. 2008).
The distribution also matches, in large part, distributions exhibited by apterous stoneflies (Plecoptera) as shown by McCulloch et al. (2010) that suggest late Pleistocene glaciation as a simultaneous vicariant event. We assume the same for A. laticorne and are of the opinion that this is a classic gap distribution with some “regional” distribution (Trewick & Wallis 2001) along mid Westland. The Southern Alps appear to have been a barrier to dispersal eastward, except for some penetration at Lake Tekapo.
Absence of A. laticorne from the North Island indicates the species did not reach Cook Strait until after it was formed some 450 000 years BP (mid Pleistocene) and did not manage to cross during the lowering of sea levels in subsequent glaciations. This is perhaps an indication that at such times climatic conditions across the Farewell Rise were not conducive to dispersal. We consider there are many suitable habitats for A. laticorne available in the North Island.
Given the occurrence of A. australense, A. stewartense, A. vailavoense, and A. ungulatum on both Stewart Island and in the south of the South Island, a not unreasonable expectation is that A. laticorne might also occur on Stewart Island; so far it is not known from there. Does this indicate that A. laticorne colonised the south of the South island after the Last Glacial Maximum when Foveaux Strait was re-submerged. As assumed for Cook Strait during glaciations, climatic extremes may well have made Foveaux Strait impassable to simuliids for much of the time. Suitable habitats for A. laticorne are not plentiful on Stewart Island; still, other species find equable localities.
Austrosimulium stewartense
The distribution of A. stewartense (Map 13) mimics, in many ways, that of the southern segregate of A. australense by occurring on Stewart Island and in the southern part of the South Island. Because of variability in the morphological features used for identifying A. stewartense and those of the closely related A. multicorne (see species descriptions), we are not confident in commenting on the identities of two apparent populations from Mid Canterbury. With a broad connection (Fig. 518) across Foveaux Strait during the Last Glacial Maximum resulting from sea level depression, and no doubt such depressions previously, the A. stewartense distribution pattern is indicated as being of Pleistocene age, but could have come about due to earlier lowered sea levels (e.g., Fig. 515). However, a more recent origin would be in agreement with the lack of molecular divergence within the tillyardianum species-group (Fig. 508a, 508b).
Austrosimulium extendorum
Austrosimulium extendorum and A. stewartense are sister species (Fig. 507) and morphologically similar. Little can be said about A. extendorum, except it is only known from Stewart Island and nearby Big South Cape Island (Map 8). With no new material, molecular data were not available, but with the lack of discrimination in the tillyardianum species-group, such material would likely have been of little use.
Austrosimulium fiordense
Dumbleton (1973) considered A. fiordense to be a subspecies of A. multicorne, but we raised it to specific status on the basis of distinctive features of the pupal thorax. Known originally only from material from Fiordland with the majority lost subsequently, new material was taken by TKC at high altitude on the Darran Mountains, Fiordland and by DAC at Rangitata River (Map 9). It is not clear if this is a gap distribution, or will be found to be a colonisation distribution (Trewick & Wallis 2001). We expect the latter, with other high altitude intermediate populations being discovered between Fiordland and the Rangitata locality (NZS126). If not, then glaciations can be invoked for the gap between populations. Irrespective, all known localities were heavily glaciated during the Otira Glaciation and current presence must be recent.
ungulatum species-group
ungulatum-subgroup
A. ungulatum, A. vexans & A. campbellense
Absent from the North Island, A. ungulatum is well distributed in the South Island and Stewart Island (Map 16). Apparent absence from parts of Marlborough is possibly a result of lack of collecting, but it could also be due to drier conditions and a lack of forest cover, which appears to be favoured by this species. Shaded cold-water streams are also lacking from much of the Canterbury Plains and Otago where it has not been collected. However, a recent collection from the Lindis Pass (NZS181, High Country ecoregion) shows that A. ungulatum does not require forest, but more obviously, cold water. As with A. multicorne, the apparent concentration of localities in the eastern foothills merely reflects accessibility from roads (Map 12). Indeed, the distribution of roads in the South Island is all too clear from the collection records for the species.
The presence of two female A. ungulatum on Banks Peninsula was unexpected — and DNA data was not recoverable from the specimen examined. Opportunities for wind dispersal from the west must be common, and more so in the past during glaciations when westerly winds are thought to have been exacerbated (Fitzharris et al. 1992) and ecological regions pushed farther east by glaciation (Burrows 2005). While present ecological conditions at the upper Kaituna River would be suitable for A. ungulatum, elsewhere it is generally not. Overall ecological conditions on Banks Peninsula have been substantially altered by human intervention particularly over the last century or so (Winterbourn 2008). Is it possible that the 2 specimens are relicts of more abundant populations present when the Peninsula was well forested?
The detailed distribution of A. ungulatum shows a small “beech gap” of less than 100 km in Westland (Map 16). The molecular data (Fig. 514) show that haplotype #38 matches closely that overall gap distribution, with haplotype #36 being more widely separated. Such distributions are normally ascribed to the glaciation that occurred in the region during the Last Glacial Maximum, with failure to close the gap since then. However, the presence of haplotype #38 at localities in the Arthurs Pass region, well glaciated during the Last Glacial Maximum, suggests that dispersal is not the problem, and that the gap is the result of other effects such as habitat availability. Further, other species of Austrosimulium do not show that gap.
The morphological cladistic analysis show A. vexans + A. campbellense as sister to A. ungulatum with A. dumbletoni sister to those (Fig 507). This result conforms to that of Dumbleton (1973) and is in agreement with his contention that the first two species were derived from A. ungulatum. Of note, though, is that molecular evidence indicates (Fig. 514) that A. vexans is sister to a small segregate (haplotype #33) of A. ungulatum, and not that species in the aggregate (Fig. 507). Divergence between the clade of A. vexans + A. ungulatum (haplotype #33) and the remainder of the ungulatum haplotypes (#35–38) is ca 7%, indicating divergence of the two clades some 3.0 Mya (early Late Pliocene). With such divergence, the vexans segregate of A. ungulatum may well be deemed a separate species if morphological characters can be determined — none can yet be found to justify this though. Is this a good example of molecular divergence without concomitant morphological change?
The main question, though, is how did the ancestor of A. vexans get to the Auckland Islands, and, for that matter, A. campbellense to Campbell Island? For the latter while morphological evidence suggests a relationship with A. vexans (Fig. 507), eventual molecular data might well reveal it to be sister to another segregate of A. ungulatum, thus representing a separate origin to that of A. vexans. We consider this worth investigating.
In the early Late Pliocene the world was emerging from a markedly warm period with elevated sea levels, and was entering, from then on, regular glaciations (Haq et al. 1987; Naish 1997). Both the Auckland Islands and Campbell Island were well emplaced and no doubt much larger having not suffered the erosion evident now. Further, hot spot islands gradually sink as the heat from their original volcanism wanes (Craig 2003), so these two islands would have then been not only larger, but higher. However, because of lack of glaciation world-wide, sea levels were higher then than now (Haq et al. 1987); hence land area would have been reduced. Naish (1997) showed regular occurrence of glaciations from 2.8 Mya (upper Late Pliocene) on, and these in general reduced sea levels between some 70–100 m. Similarly, apart from the Otira and Waimea Glaciations when sea levels plummeted to possibly -135 m, over the last 250 000 years the usual depression during cold periods, has been more of the order of 70 m (Fig. 515) (e.g., Chappel 1986; Burrows 2005). So, exposed land south of Stewart Island and north of Auckland Islands during these periods, although not as extensive as that in the Last Glacial Maximum (Fig. 518), would still have reduced the distance for dispersal of an A. vexans precursor, and similarly for the precursor to A. campbellense to Campbell Island.
Dispersal by wind is a distinct possibility. McGlone (2002) had substantial evidence for long-distance pollen dispersal (pollen rain) to New Zealand’s Southern Oceanic Islands, with New Zealand the dominant source. However, he also noted that the Auckland Islands show abundant evidence of glacial activity. The U-shaped valley systems of Campbell Island also indicate glaciation, though more ambiguously.
McGlone (2002) also considered problems that during recent glaciations, the marine Antarctic Convergence possibility moved 5 degrees north. This would have likely resulted in Auckland and Campbell Islands possessing climates similar to that now found on Macquarie Island; that is, summer temperatures as low as 5 to 6°C and 3 to 4°C, respectively. That would have created problems for the present flora, since cooling would have resulted in no suitable land for survival because of the depression of snow level. Not mentioned, however, is that with the major glaciation during those recent glaciations, sea level depression was marked, even down to 135 m below present levels (Fig. 518). So, while the current land areas of Campbell and Auckland Islands are 113 km2 and 626 km2, respectively, during the Last Glacial Maximum they would have been some 4 000 km2 and 6 800 km2, respectively, for perhaps 5 000–10 000 years during each of the Otira and Waimea Glaciations (Fig. 515).
Opportunities for colonisation of Auckland and Campbell Islands by Austrosimulium were not lacking — running water would not have been in short supply. The molecular data indicate early Late Pliocene; the probability is low that colonisation of Auckland and Campbell Islands by Austrosimulium was older. While the islands are some 19–12 and 11–6 million years old respectively (mid Early Miocene), the Campbell Plateau, which might have served as a biological source, was well submerged by 40–35 Mya (Middle Eocene) (Sutherland et al. 2010: their Fig 2). That is, there was a gap of, at maximum, 16 million years between the flooding of the plateau and formation of land via volcanoes. This presents a major problem for the origin of all the biota of these islands, some of which is thought to be old in nature (e.g., Michaux & Leschen 2005).
Dumbleton (1963b) was of the opinion that A. vexans was derived from A. ungulatum via aerial colonists, post-Pleistocene, but he did not elaborate. In 1973, he commented that origin of A. campbellense was probably similar to that of A. vexans and post-Pleistocene, at oldest post Pliocene. We agree.
The remaining moderately well supported clade of the ungulatum group s.s. (haplotypes #35–38) has haplotype #35 (NZS82, Lyell Creek, Buller River), sister to the remainder at 1.9% divergence; indicating the age of separation at maximum 826 000 year BP (early Late Pleistocene). This location is of interest in that the general region is considered to have harboured relictual forest during the Last Glacial Maximum (Alloway et al. 2007) and no doubt during others previously. Austrosimulium ungulatum appears to be dependent on forest, although the exact reason why is not obvious. Small, cold-water, densely-shaded streams are normal for these larvae, although shade is not necessary (e.g. NZS181). Could this region be the origin of A. ungulatum s.s?
The remaining A. ungulatum haplotypes (#36–38) had a divergence of some 3.7 % from the others, which gives an origin at ca 1.6 Mya (upper Lower Pleistocene). In general, the haplotypes are widely dispersed and their apparent absence, in particular from Kaikoura, Canterbury, and Otago regions, are sampling artifacts (see Map 16). There is little in the northern distribution to indicate any effect of the Southern Alps as a distributional barrier, as suggested for other invertebrates (McCulloch et al. 2010). Although molecular data were not available for the localities at Kaikoura (Map 16), we expect them to be probably of haplotype #38, distributed to the west.
Still, an apparent “beech gap” for A. ungulatum is puzzling, as is its absence from the North Island. Females of A. ungulatum will fly considerable distances to blood feed. They develop eggs and need to find running water to oviposit. Running water is not lacking in the region of the “beech gap”, so why has A. ungulatum not filled that gap? Other previously glaciated areas have A. ungulatum. A partial reason may involve the chemistry of water in the region, much of which is brown from dissolved organic compounds. Austrosimulium larvae appear to be tolerant of a wide range of pH, and elevated acidity due to organic compounds does not alter presence of many invertebrate taxa (Winterbourn & Collier 1987). However, acid streams, in comparison to those that are alkaline, may have only a quarter of the density of aquatic invertebrates (Collier & Winterbourn 1987) and that too was our experience with aquatic macroinvertebrates. So, although presence of the “beech gap”, is assumed to be due to glacial effects, is it really a consequence of ecological conditions, or a combination of both? That haplotype #38, which occurs on either side of the apparent gap is also found at Arthurs Pass, which recently glaciated, points to ecological reasons for the gap to the west.
Relevant to the ecological reasoning for a beech gap is the rapid fall-off in abundance of Austrosimulium larvae south of the Kawhaka River, Old Christchurch Road, a light brown-water river (NZS50), near the northern edge of the “beech gap”. Yet, a mere 27 km directly ENE, is currently New Zealand’s most species-rich simuliid locality at Jacksons Bridge, Taramakau River (NZS51). There, larvae of A. australense, A. laticorne, A. multicorne, and A. tillyardianum, together with adults of A. ungulatum were obtained from a small area. This is unusual even though the Taramakau River has clear water, but it is a large, braided, unstable river and not obviously suited to all those species. Still, A. australense (Map 3) and A. multicorne (Map 12) occur sporadically along the gap region.
The apparent aggregation of haplotypes in the Arthurs Pass region (Fig. 514) is a sampling artifact, as material from several nearby localities (Map 16) was not available for molecular analysis.
The two main haplotypes of A. ungulatum (#36, #38) are widespread and overlap in distribution. Their divergence of 2.5% represents separation of lineages at about 1 Mya (upper Lower Pleistocene). At that time New Zealand was emerging from a period of some 20 cool episodes that alternated with warm periods before entering more substantial glaciations (Burrows 2005). It was also the time of the devastating eruption that covered much of the North Island and produced the Kidnappers Ignimbrite. Is it possible that there was a genetic bottleneck at that time for A. ungulatum with subsequent widespread dispersal of the two haplotypes over the South Island? And attesting to the apparent considerable flight ability of the females?
Dumbleton (1973), too, puzzled over the distribution of A. ungulatum, in particular its absence from the North Island. For the adults, some dispersal and aggregation behaviour is well known. Part of the puzzle surrounding the absence of A. ungulatum in North Island is that Mount Taranaki currently has conditions that appear to be suitable for this species. But that was not so during the Last Glacial Maximum. As McGlone & Neall (1994), Alloway et al. (2007), and others have shown, the mountain was then covered in shrubland and grasses that do not provide optimal habitat for A. ungulatum. Further, the mountain is relatively young (0.015–0.012 Mya) and has been volcanically active. The change from shrubland to forest was rapid, possibly over a period of only 1 000 years around 12 000 years BP (McGlone & Neall 1994). Furthermore, conditions for an adult simuliid attempting to fly across the present Cook Strait are not amicable, and strong westerly winds that are typical will reduce the probability of dispersal from the South Island to the North. Such winds were known to have been exacerbated during glaciations. Still, such winds do not blow continuously.
A not unreasonable expectation would be for Austrosimulium and other aquatic insects to disperse easily between the two main islands during glacial maxima because of rivers flowing out onto the South Island–Taranaki plain. The Waimea, Motueka, Takaka, and Aorere Rivers no doubt flowed out onto the plain from the South Island, as did the Whanganui and the Rangitikei Rivers from the North Island. The exact drainage patterns are not known, but some have been suggested (Fleming 1975; Lewis et al. 1994; Trewick & Bland 2011). Even if the rivers did not merge, distances between them would have been short.
unicorne-subgroup
Austrosimulium vailavoense + A. unicorne & A. bicorne + A. tonnoiri + A. dumbletoni
Dumbleton (1973) considered A. unicorne and A. bicorne to constitute taxonomically the unicorne-subgroup, sister to the New Zealand ungulatum-subgroup. Our cladistic analysis is in full agreement at the Majority Rule level (Fig. 506), with minor differences when fully resolved (Fig. 507). We raised A. tonnoiri to species status from specimens considered by Dumbleton to be A. bicorne, mainly on the basis of the distinct pupal cocoon. Both A. dumbletoni and A. vailavoense are only known from adults, and morphologically appear to belong to the ungulatum-subgroup.
Molecular analysis (Fig. 514) is slightly at variance to the above, with A. vailavoense and A. dumbletoni clustered in what is normally considered the unicorne-subgroup and discussion below is centred round this point. The placement of these two species in the molecular phylogeny is perhaps not anomalous given the lack of morphological divergence in the females of the ungulatum species-group. Species designation is in large part based on pupal horn and larval anal sclerite characters, however, immatures of the two species of concern are not yet known. If the placement of these two species based on molecular data is correct, then we suggest that gills of pupal A. vailavoense and A. dumbletoni, when discovered, will be more similar to those of the unicorne species-group, with a basal horn and fine filaments (Fig. 283, 285), rather than the antlered type found in ungulatum species s.l. (Fig. 281).
The unicorne-subgroup is part of New Zealand’s unusual high-alpine biota which has long been of interest (e.g., Fleming 1963, 1979; Johns 1969; Raven 1973; McGlone 1985; Given & Gray 1986; Morgan-Richards & Gibbs 1996; Trewick et al. 2000; Hitchings 2009). In large part the interest is because of the relatively youthful age of the Southern Alps.
There is ample tectonic evidence (e.g., King 2000; Coates 2002) to suggest that the Southern Alps are at the most perhaps 8 million years old, but more likely 5 or even 3, when uplift was accelerated to produce their present form. Batt et al. (2000) used K–Ar dating to determine that their age was certainly in the order of 5 million years. Of note, they commented that estimates of earlier age (e.g., 8 Mya) were based on older rock in Fiordland exposed during later exhumation erosion along the mountains, a process recently examined in detail by Shuster et al. (2011). Chamberlain et al. (1999) and Chamberlain & Poage (2000) using d18O isotope also showed changes that agreed with an origin of the Alps at 5 Mya and, further, that there was a rapid 2 000 m increase in altitude at that time in the early Miocene.
Is this an acceptable age for the origin of Austrosimulium in New Zealand, via dispersal from Australia, probably Tasmania? Certainly, sister taxa in Tasmania are adapted to cool temperate conditions, as are those of South America (Fig 507). Preadaptation to mountainous conditions in New Zealand has been suggested already by Dumbleton (1973). Molecular evidence from divergence ages of lineages would suggest so and is not at variance with that of paleogeology.
A nagging question remains, though, in regards to the other high alpine aquatic fauna such as blepharicerids and mayflies, both considered to be older groups. Did they, too, adapt to the alpine conditions at the same time as Austrosimulium, or were there separate events?
The newly recognised species A. vailavoense from the southern South Island and Stewart Island (Map 18) shows a distribution common to southern segregates of other species, namely A. australense, A ungulatum, and A. stewartense. While we did not include A. vailavoense in the cladistic analysis because we lack all stages other than female adults, we currently place it in the ungulatum-subgroup following the same reasoning as Crosby (1976a) for A. dumbletoni — namely the toothed claw.
A marked molecular divergence of ca 16% from the ungulatum-subgroup proper (Fig. 513), indicates that the precursor to A. vailavoense was an older entity, perhaps up to some 7 million years old. The two populations of this species (NZS157, 165) separated by Foveaux Strait are divergent at only 0.7%, indicative of more recent dissociation, some 330 000 years BP (mid Middle Pleistocene), perhaps following the end of the Nemona Glaciation (Burrows 2005) when sea level depression is estimated to have reached at least -110 m and then risen rapidly (Huybrechts 2002). The older divergence from the common ancestor with A. unicorne presents something of a problem, since that species is now strictly a cold-water, high-altitude species. These are habitat attributes that were not available until the Early Pliocene (5 Mya) with orogeny of the Southern Alps.
Based on the slightly longer divergence of the South Island population of A. vailavoense one scenario is that the precursor of A. vailavoense diverged from that of A. unicorne on the South Island and then dispersed to Stewart Island. Austrosimulium unicorne then adapted to alpine conditions when these became available later, with A. vailavoense inhabiting lowland forested habitats. This latter assertion will be tested when early stages of A. vailavoense are discovered.
For A. unicorne, only a single locality (NZS132, Pegleg Creek, Arthurs Pass) was available for molecular analysis. Other localities (Map 17) show this species to be restricted to the Southern Alps in the central South Island. With one exception, all localities have cold, clear water at high altitude. Molecular data from other populations should reveal how divergent they are — if at all. At present the distribution fits the “colonisation” definition of Trewick & Wallis (2001) consistent with reinvasion of suitable habitats following the Last Glacial Maximum.
Austrosimulium unicorne (Fig. 514), as sister to A. vailavoense, is divergent at ca 9% — indicating a split of some 4 Mya, and an age similar to that of the Southern Alps.
The sister clade to the above is comprised of A. bicorne, A tonnoiri, and A. dumbletoni. With the exception of A. dumbletoni, for which there is not much information, the other two have similar ecological requirements to A. unicorne (Fig. 507). Divergence time for this clade from the ungulatum species-group is similar to that for its sister clade (A. vailavoense + A. unicorne), and is relatively consistent with formation of the Southern Alps. Austrosimulium bicorne is known from two regions, one (haplotype #41) in the Arthurs Pass region (NXS133) and the other (haplotypes #42, 43) in Fiordland (NZS32) (Map 4). The haplotype #41 from Arthurs Pass is divergent from haplotype #42 by some 2.8%, an age of some 1.2 million years. These two populations should be re-examined morphologically for the possibility that they are cryptic species. The Fiordland population (NZS32) shows evidence for perhaps 3 haplotypes, with #42 divergent at 1.2 % (522 000 years BP) and another at ca 1.0 % (435 000 years BP). Is this indicative of multiple isolation events via glaciations? Their occupation of areas that were heavily glaciated during the Otira Glaciation, indicates an ability to disperse rapidly.
For the present we consider the gap between Arthurs Pass and Fiordland to be a collecting artifact and expect other intermediate populations to be discovered. Collecting this species and A. tonnoiri, elsewhere in the Southern Alps will, however, require considerable physical stamina. Mountaineering skills would not be amiss.
Austrosimulium unicorne, A. bicorne, and A. tonnoiri are all high altitude and cold water species. In addition, they have further markedly specialised requirements for the larvae. These immatures are large and pale. UV radiation is avoided by simuliid larvae (Donahue & Schindler 1998; Kelly et al. 2003), so it is likely that their usual habitat (perched stones) allows larvae to attach on the undersides, thereby avoiding UV radiation while still in full flow of water. Further, in the alpine streams inhabited by these larvae the water levels vary considerably, often on a diurnal basis as snow packs melt. Dumbleton (1973: 545) suggested that the thick pupal cocoons of these species (Fig. 175–177) would protect against temperature fluctuations; an unusual suggestion for a poikilotherm such as a simuliid. We consider that it is far more likely to be protection against desiccation from the rapidly varying water levels. Such distinctly specialised habitats cannot be older than the age of the Southern Alps and specifically require mountainous terrain to form (Frey & Church 2009). This represents physical evidence that origin of these species is no more than 5 Mya.
Known only from Westland at Jackson Bay and Knights Point some 60 km further north (Map 7), A. dumbletoni appears to have a refugial distribution. Burrows (2005, his Fig. 7.1) showed both areas as being non-glaciated during the Otira Glaciation, but ice extended out to sea between them. Collection efforts in similar ice-free areas farther north would test this assertion.